Kitz ADSL Broadband Information
adsl spacer  
Support this site
Home Broadband ISPs Tech Routers Wiki Forum
 
     
   Compare ISP   Rate your ISP
   Glossary   Glossary
 
Please login or register.

Login with username, password and session length
Advanced search  

News:

Pages: 1 ... 14 15 [16] 17 18 ... 20

Author Topic: G.INP enabled Stats Comparison  (Read 98768 times)

Chrysalis

  • Content Team
  • Addicted Kitizen
  • *
  • Posts: 7405
  • VM Gig1 - AAISP CF
Re: G.INP enabled Stats Comparison
« Reply #225 on: April 02, 2015, 09:51:04 PM »

those stats are g.inp only?
Logged

Bald_Eagle1

  • Helpful
  • Kitizen
  • *
  • Posts: 2721
Re: G.INP enabled Stats Comparison
« Reply #226 on: April 02, 2015, 10:05:11 PM »

I do log 24/7 using Dslstats on the RPi it's just the HG612_Modem_stats uses MS windows OS for the program to work and the PC go's to sleep the same time as me ;)


I'd also forgotten that too.

I think I should get my coat now!  ::)
Logged

WWWombat

  • Kitizen
  • ****
  • Posts: 1674
Re: G.INP enabled Stats Comparison
« Reply #227 on: April 03, 2015, 12:14:25 AM »

Here are my equivalent graphs.

Original G.INP activation 27/03/15 at 06:00ish
Manual resync just before 28/03/15 00:00
Logged

NewtronStar

  • Kitizen
  • ****
  • Posts: 4898
Re: G.INP enabled Stats Comparison
« Reply #228 on: April 03, 2015, 12:37:34 AM »

It would help if i knew what LEFT meens and I can guess the RS is re-transmit ;)
Logged

WWWombat

  • Kitizen
  • ****
  • Posts: 1674
Re: G.INP enabled Stats Comparison
« Reply #229 on: April 03, 2015, 01:16:44 AM »

I think:

EFTR = Error-Free Throughput

Calculated once per second and, I guess, depends on how many packets need to be retransmitted.

LEFTR = A Low EFTR defect

This defect is an event (a bit like discovering a CRC error, or a FEC) that gets logged whenever the calculated EFTR goes below a threshold value (defined by LEFTR_THRESH).

LEFTRS = A counter of second-long periods where one or more LEFTR defects were logged. In this sense, it works just like the ES counter does, as a time-based monitor for the number of CRC errors.

I'm pretty sure the first two definitions are right, but the spec doesn't use the term "LEFTRS". Instead, it always goes longhand with "leftr defect seconds counter". However, it seems to be the best choice.

What does it mean? With CRC's being a thing of the past (anything which used to get a CRC error will now get re-transmitted), I think LEFTRS will be used as the equivalent of ES.

There is also a "severe low EFTR defect" known as SEFTR, but there doesn't appear to be a SEFTRS counter; I *think* SEFTR events get counted in the existing SES counter.
Logged

Bald_Eagle1

  • Helpful
  • Kitizen
  • *
  • Posts: 2721
Re: G.INP enabled Stats Comparison
« Reply #230 on: April 03, 2015, 07:22:48 AM »

Here are my equivalent graphs.



Which, if I have correctly understood the various articles I have read recently, along with your own comments, suggest/confirm that your connection is hardly affected by crosstalk (or other 'interference') at all.


Logged

hadookaan

  • Just arrived
  • *
  • Posts: 8
Re: G.INP enabled Stats Comparison
« Reply #231 on: April 03, 2015, 03:09:08 PM »

Hi all I used to get a ping of 4-5ms and just noticed its gone up quite alot to 15-16ms is this due to inp, and is there any way i can revert back to my old ping...as all this nonsense is annoying me!!

this is my connection statistic if that helps!
39952 kbps   10000 kbps
Logged

lloyd

  • Reg Member
  • ***
  • Posts: 109
Re: G.INP enabled Stats Comparison
« Reply #232 on: April 03, 2015, 03:29:26 PM »

Hi all I used to get a ping of 4-5ms and just noticed its gone up quite alot to 15-16ms is this due to inp, and is there any way i can revert back to my old ping...as all this nonsense is annoying me!!

this is my connection statistic if that helps!
39952 kbps   10000 kbps

probably best if you ask a question in one place. Some one has already responded to your duplicate question on another thread
Logged

roseway

  • Administrator
  • Senior Kitizen
  • *
  • Posts: 43603
  • Penguins CAN fly
    • DSLstats
Re: G.INP enabled Stats Comparison
« Reply #233 on: April 03, 2015, 03:58:03 PM »

Agreed. I've moved hadookan's query to a separate thread here: http://forum.kitz.co.uk/index.php/topic,15273.msg283896.html#msg283896
Logged
  Eric

WWWombat

  • Kitizen
  • ****
  • Posts: 1674
Re: G.INP enabled Stats Comparison
« Reply #234 on: April 03, 2015, 04:52:03 PM »

Which, if I have correctly understood the various articles I have read recently, along with your own comments, suggest/confirm that your connection is hardly affected by crosstalk (or other 'interference') at all.

I don't know if it really does show crosstalk, or just noise of any source. Nevertheless, I only used to get around 50 ES's per day, from maybe 80 CRC's, so I wasn't heavily affected by anything anyway. I do however, see odd increases and decreases in SNRM that look awfully like someone, somewhere turning a modem on or off; SNRM gets affected then, but error rate doesn't change.

G.INP activation came with a small amount of FEC and interleaving on bearer 0; DSLstats tells me that I'm now averaging maybe 1,000 FEC's per day. It looks like this is mopping up the errors that my line used to suffer from. I see no RSuncorr now, which, on the face of it, means I shouldn't be troubling the retransmission portion of G.INP.

However, the retransmission counters are increasing, so it looks like *some* packets get retransmitted. I guess LEFTRS gets incremented, not just when a packet gets retransmitted, but when "a lot" of packets need retransmitting at roughly the same time - so the graphs suggest I've had a couple of those occasions. The absence of RSuncorrs therefore seems to be a red herring.

Can you post up your sets of graphs for, say, the last 10 days?
Logged

NewtronStar

  • Kitizen
  • ****
  • Posts: 4898
Re: G.INP enabled Stats Comparison
« Reply #235 on: April 03, 2015, 09:24:36 PM »

I don't know if it really does show crosstalk, or just noise of any source.

crosstalk will manifest itself as noise into the effected end-users line but the difference is it's another broadband signal thats interfering with the EU's line rather than the normal culprits like RFI and REIN this is were vectoring comes in it's able to mask out the crosstalk/noise from the disturbing broadband signal.
« Last Edit: April 03, 2015, 09:44:41 PM by NewtronStar »
Logged

Bowdon

  • Content Team
  • Kitizen
  • *
  • Posts: 2395
Re: G.INP enabled Stats Comparison
« Reply #236 on: April 03, 2015, 11:35:42 PM »

I had a unique opportunity last night to see what happened with my modem stats when there was a power cut across 99% of the town. Luckily my house/modem kept on, and so did the cabinet.

I noticed that my SNR numbers increased and I got very few errors, no error seconds at all. I actually thought the program had bugged out as nothing was moving lol.

It makes me wonder if the phone cables from the cabinet are near to power lines.

I'm hoping g.inp will counter this low level interference when it arrives to my eci cabinet.
Logged
BT Full Fibre 500 - Smart Hub 2

NewtronStar

  • Kitizen
  • ****
  • Posts: 4898
Re: G.INP enabled Stats Comparison
« Reply #237 on: April 04, 2015, 01:30:57 AM »

I'm hoping g.inp will counter this low level interference when it arrives to my eci cabinet.

It will if G.INP works the same way as it does on the Huawei cabinet, it does a good job on REIN interference and RFI and SHINE :yay:
« Last Edit: April 04, 2015, 01:40:13 AM by NewtronStar »
Logged

Bald_Eagle1

  • Helpful
  • Kitizen
  • *
  • Posts: 2721
Re: G.INP enabled Stats Comparison
« Reply #238 on: April 04, 2015, 06:19:20 AM »

Can you post up your sets of graphs for, say, the last 10 days?


I'm away from home at the moment, so I'll post the graphs on my return.

In the meantime, my Bearer 0 data graphs can be viewed on MDWS.



Pre-G.INP, I was on fastpath, but did see low numbers of DS RSCorr/FEC, RSUnCorr, CRC & HEC counts.
ES were generally at around 1000 per day.

ES CRC, HEC & RSUnCorr have now all but disappeared, but I now see fairly constant (& lowish?) levels of rtx_tx, rtx_c & an occasional rtx_uc.
I also now see quite a few RSCorr on Bearer 0 & low levels on Bearer 1.

Logged

lockeh

  • Member
  • **
  • Posts: 21
Re: G.INP enabled Stats Comparison
« Reply #239 on: April 04, 2015, 10:58:40 AM »



Anyone know why my line has interleave enabled?

before the rubbish G.INP was forced on my line I had 3 years worth of full sync fast path trouble free line.

The line is only about 80 meters long of underground .5 copper.
Logged
Pages: 1 ... 14 15 [16] 17 18 ... 20